The primary aim of this research was to determine whether teachers in their initial or in-service training have acquired a basic knowledge of the interpretation of children's drawings of human figure. A literature study was done in which the focus was placed on major aspects such as: • The normal development of children's drawings according to different stages, and • Those drawings that deviate from the normal development and their implications. • Reliability of children's drawings as a projective tool. A measuring instrument was developed in order to measure the basic knowledge of teachers with regard to the interpretation of children's drawings of human figure. The results of the empirical research indicated that teachers have limited basic knowledge with regard to the interpretation of children's drawings. The educational implications with reference to teachers, education system and other professionals were discussed. Recommendations for further research on similar topics were made. / Educational Studies / M. Ed. (with specialisation in Guidance and Counselling)

As theories on the process of reading have advanced so definitional changes have resulted. This in turn has impacted on the teaching of reading. Comprehension is not the result of successful word recognition, rather, meaning is constructed by the reader using various sources as a frame of reference. Reading requires an interaction between the reader, the text read and the context in vvhich it is read. Reader factors involve language competencies, prior knowledge, vocabulary, the use of strategies and attitudes and motivation. The text may be narrative or expository and encompasses instructional materials. The tvvo broad categories in the context are the classroom setting and the instructional context. An 'interactive' or 'organisational' paradigm underlying inclusion recognises individual differences as being a probable cause of failure but postulates that the school and all that it encompasses, can be a barrier to learning and development. As such it is not deficit driven, attributing failure to learners alone. In the case of reading, this means not attributing reading failure to the reader alone but acknowledging the role of the text and the context. This point of departure is confirmed by an interactive model of disability, which, while still explaining reading deficits, advances that alternative areas also be investigated. Focusing on abilities is conducive to proactivity in the prevention of barriers to learning and development. To be inclusive therefore, mainstream schools generally and classrooms specifically, will need to be reformed and restructured to be more responsive to learners experiencing barriers to learning and development. This will require enhanced teaching methods and flexible support systems. Accommodating diversity presupposes the acknowledgement of each learner's uniqueness in order to meet individual needs. This will be facilitated when in the compilation of a reading programme to meet individual needs, the reader, the text and context are matched through assessment and instruction . A reading programme to enhance the teaching of reading and thereby meet individual needs has been compiled and implemented in a mainstream, intermediate phase class. From this practical experience and the literature studied, guidelines for the teaching of reading have been formulated for teachers in the intermediate phase to use within the context of inclusion. / Educational Studies / D. Ed. (Orthopedagogics)

Many researchers (e.g., Nation, 2001, 2015; Schmitt, 2000) have recognized the importance of vocabulary learning in second language (L2) or additional language (AL) acquisition. The strong effects of lexical and background knowledge on L2reading comprehension have similarly been found in various studies (e.g., Hu & Nation, 2000; Rokni & Hajilari, 2013). In the case of Japanese language, the opportunities for acquiring the lexical and background knowledge associated with Japanese history and culture, especially traditional culture, are scant, because only a small number of Japanese pedagogical materials deal minimally with these topics. Meanwhile, many learners are motivated to study Japanese because of their interest in Japanese history and culture, according to a survey conducted by the Japan Foundation in 2012. This project aimed to increase the opportunities for learning Japanese history and traditional culture through the development of new pedagogical materials based on kabuki, and then the empirical evaluation of the developed pedagogical materials. Nine Chinese-as-a-first-language Japanese learners at the upper-intermediate level participated in the nine-week online course, including the pre- and post-course tests in the first and last weeks. Employing a multi-method research approach, the study examined the changes in learners’ lexical and background knowledge related to Japanese history and culture, their reading comprehension, and their interest in kabuki. Four kinds of multiple-choice tests were administered to collect the quantitative data. In addition, the qualitative data were gathered through the pre- and post-course questionnaires and post-course individual interviews. Overall, the findings indicated that almost all participants increased their background knowledge of kabuki, as well as their vocabulary related to kabuki and general theatrical performances. The results in other areas, such as historical vocabulary, vocabulary depth, reading comprehension, and historical background knowledge were mixed. Further, concerning the depth of vocabulary knowledge, it was found that the learning of vocabulary depth was more difficult than learning of vocabulary breadth. Likewise, the knowledge of use, such as collocations and register constraints, was found to be more difficult to learn than other aspects of vocabulary depth. The participants’ reports in the post-course questionnaire and individual interviews showed that most participants seemed to have increased their interest in kabuki. Overall, the first-of-their-kind developed pedagogical materials contributed to the development of lexical and background knowledge, specifically knowledge associated with Japanese traditional culture and history. This study may provide a model for an evidence-based approach to the development of pedagogical materials that practitioners can adopt or adapt. / Graduate

Abstracts in English, Afrikaans and Zulu / The purpose of this study was to explore the learners’ understanding of trigonometric functions using GeoGebra software. A qualitative case study approach was used with six Grade 11 learners at a school in Tshwane South District. The data were collected during a seven-day period using multiple methods: a diagnostic test, worksheets, a smart recorder, a trigonometric functions test, one-on-one interviews and focus-group discussions. The findings showed clearly that the use of GeoGebra enhanced Grade 11 learners’ understanding of trigonometric functions. The findings also showed that the use of GeoGebra helped the learners to understand the basic trigonometric functions graphs. This enabled them to sketch freely without using the point-by-point method. Based on this study, it is recommended that GeoGebra should be made available to all Grade 11 learners. This will encourage them use the software out of regular school hours. / Die doel van hierdie studiegids was om die leerder se kennis van die gebruik van GeoGebra sagteware, ten opsigte van trigonometriese funksies te ondersoek. ‘n Kwalitatiewe benadering is gevolg met ses Graad 11 leerders by ‘n skool in die Tshwane Suid Distrik. Die data is ingewin oor ‘n periode van sewe dae, deur gebruik te maak van verskeie metodes: ‘n diagnostiese toets, werkkaarte, ‘n “smart” opname, ‘n trigonometriese funksies toets, een-tot-een onderhoude en fokusgroepe waar besprekings plaasgevind het. Die data-analise wys duidelik dat die gebruik van GeoGebra, die Graad 11 leerders gehelp het om trigonometriese funksies beter te verstaan. Die uitkomste wys ook duidelik dat GeoGebra die leerders gehelp het met ‘n beter begrip van die basiese trigonometriese funksies. Dit het hulle gehelp om vryhand sketse te doen en nie noodwendig die punt-tot-punt metode nie. Gebaseer op hierdie studie beveel ons aan dat GeoGebra beskikbaar gemaak moet word aan aale Graad 11 leerders. Dit sal leerders motiveer om ook die sagteware op hulle eie buite skoolure te gebruik. / Injongo yocwaningo lolu bekuyikuhlola ukuzwisisa kwabantwana amagrafu e-trigonometry ma bewafunda nge softhiwe ye GeoGebra. Abantwana abayisithupha abakubanga letshumi lanye abaphuma kusigodi se Tshwane South babambiqhaza kulolucwaningo lwendlela ye ‘qualitative case study’. Imininingo iqoqwe kumalanga ayisikhombisa kusetshenziswa indlela ezilandelayo: ukuhlolwa kwe-diagnostic, amaphepha okusebenzela, isingxoxo zamunye ngamunye lezingxoxo leqembu. Iziphumo ezinkulu zikhombe ngokucacile ukuthi ukusetshenziswa kwesofthiwe yeGeoGebra kukhulise ukuzwisisa kwamagrafu e-trigonometry ngabafundi bebanga letshumi lanye.Iziphumo njalo zibonise ukuthi abafundi bazwisisa izinto eziyisiseko ngala magrafu ma bewafunda ngesofthiwe ye GeoGebra. Lokhu kunike abafundi amandla okudwebadweba lamagrafu ngokushesha ngokukhululeka. Kusekelwa ngalezi ziphumo, kunconyiwe ukuthi abafundi bafumane isofthiwe yeGeoGebra ngaso sonke isikhathi. Lokhu kuzabakhuthaza ukuthi basebenzise lesofthiwe noma bengaphandle kwesikolo. / Mathematics Education / M. Ed. (Mathematics Education)

Les modèles de réseaux de neurones et en particulier les modèles profonds sont aujourd'hui l'un des modèles à l'état de l'art en apprentissage automatique et ses applications. Les réseaux de neurones profonds récents possèdent de nombreuses couches cachées ce qui augmente significativement le nombre total de paramètres. L'apprentissage de ce genre de modèles nécessite donc un grand nombre d'exemples étiquetés, qui ne sont pas toujours disponibles en pratique. Le sur-apprentissage est un des problèmes fondamentaux des réseaux de neurones, qui se produit lorsque le modèle apprend par coeur les données d'apprentissage, menant à des difficultés à généraliser sur de nouvelles données. Le problème du sur-apprentissage des réseaux de neurones est le thème principal abordé dans cette thèse. Dans la littérature, plusieurs solutions ont été proposées pour remédier à ce problème, tels que l'augmentation de données, l'arrêt prématuré de l'apprentissage ("early stopping"), ou encore des techniques plus spécifiques aux réseaux de neurones comme le "dropout" ou la "batch normalization". Dans cette thèse, nous abordons le sur-apprentissage des réseaux de neurones profonds sous l'angle de l'apprentissage de représentations, en considérant l'apprentissage avec peu de données. Pour aboutir à cet objectif, nous avons proposé trois différentes contributions. La première contribution, présentée dans le chapitre 2, concerne les problèmes à sorties structurées dans lesquels les variables de sortie sont à grande dimension et sont généralement liées par des relations structurelles. Notre proposition vise à exploiter ces relations structurelles en les apprenant de manière non-supervisée avec des autoencodeurs. Nous avons validé notre approche sur un problème de régression multiple appliquée à la détection de points d'intérêt dans des images de visages. Notre approche a montré une accélération de l'apprentissage des réseaux et une amélioration de leur généralisation. La deuxième contribution, présentée dans le chapitre 3, exploite la connaissance a priori sur les représentations à l'intérieur des couches cachées dans le cadre d'une tâche de classification. Cet à priori est basé sur la simple idée que les exemples d'une même classe doivent avoir la même représentation interne. Nous avons formalisé cet à priori sous la forme d'une pénalité que nous avons rajoutée à la fonction de perte. Des expérimentations empiriques sur la base MNIST et ses variantes ont montré des améliorations dans la généralisation des réseaux de neurones, particulièrement dans le cas où peu de données d'apprentissage sont utilisées. Notre troisième et dernière contribution, présentée dans le chapitre 4, montre l'intérêt du transfert d'apprentissage ("transfer learning") dans des applications dans lesquelles peu de données d'apprentissage sont disponibles. L'idée principale consiste à pré-apprendre les filtres d'un réseau à convolution sur une tâche source avec une grande base de données (ImageNet par exemple), pour les insérer par la suite dans un nouveau réseau sur la tâche cible. Dans le cadre d'une collaboration avec le centre de lutte contre le cancer "Henri Becquerel de Rouen", nous avons construit un système automatique basé sur ce type de transfert d'apprentissage pour une application médicale où l'on dispose d’un faible jeu de données étiquetées. Dans cette application, la tâche consiste à localiser la troisième vertèbre lombaire dans un examen de type scanner. L’utilisation du transfert d’apprentissage ainsi que de prétraitements et de post traitements adaptés a permis d’obtenir des bons résultats, autorisant la mise en oeuvre du modèle en routine clinique. / Neural network models and deep models are one of the leading and state of the art models in machine learning. They have been applied in many different domains. Most successful deep neural models are the ones with many layers which highly increases their number of parameters. Training such models requires a large number of training samples which is not always available. One of the fundamental issues in neural networks is overfitting which is the issue tackled in this thesis. Such problem often occurs when the training of large models is performed using few training samples. Many approaches have been proposed to prevent the network from overfitting and improve its generalization performance such as data augmentation, early stopping, parameters sharing, unsupervised learning, dropout, batch normalization, etc. In this thesis, we tackle the neural network overfitting issue from a representation learning perspective by considering the situation where few training samples are available which is the case of many real world applications. We propose three contributions. The first one presented in chapter 2 is dedicated to dealing with structured output problems to perform multivariate regression when the output variable y contains structural dependencies between its components. Our proposal aims mainly at exploiting these dependencies by learning them in an unsupervised way. Validated on a facial landmark detection problem, learning the structure of the output data has shown to improve the network generalization and speedup its training. The second contribution described in chapter 3 deals with the classification task where we propose to exploit prior knowledge about the internal representation of the hidden layers in neural networks. This prior is based on the idea that samples within the same class should have the same internal representation. We formulate this prior as a penalty that we add to the training cost to be minimized. Empirical experiments over MNIST and its variants showed an improvement of the network generalization when using only few training samples. Our last contribution presented in chapter 4 showed the interest of transfer learning in applications where only few samples are available. The idea consists in re-using the filters of pre-trained convolutional networks that have been trained on large datasets such as ImageNet. Such pre-trained filters are plugged into a new convolutional network with new dense layers. Then, the whole network is trained over a new task. In this contribution, we provide an automatic system based on such learning scheme with an application to medical domain. In this application, the task consists in localizing the third lumbar vertebra in a 3D CT scan. A pre-processing of the 3D CT scan to obtain a 2D representation and a post-processing to refine the decision are included in the proposed system. This work has been done in collaboration with the clinic "Rouen Henri Becquerel Center" who provided us with data

The study’s purpose was to explore the components of pedagogical content knowledge (PCK) for teaching quadratic function in Grade 10 mathematics classrooms in Mogalakwena district, Limpopo. The study was mainly guided by Shulman’s PCK and it was categorised as content knowledge, knowledge of the curriculum, knowledge of learners and knowledge of teaching strategies. The researcher employed a qualitative case study research design to explore the teaching of three purposefully selected Grade 10 mathematics teachers regarding their knowledge of quadratic function, strategies employed to teach quadratic function and knowledge of Grade 10 mathematics learners in quadratic function. In this study the researcher argued that PCK strands are interwoven, and in order to teach mathematical concepts, such as quadratic function, effectively teachers should employ components of PCK to complement one another. Data were gathered by lesson observation and interviews. Findings revealed that concepts of quadratic function are inefficiently addressed in Grade 10 due to teachers’ lack or inadequacy in some aspects of PCK. Therefore, there is need to develop mathematics teachers’ PCK in the Mogalakwena district to enhance their teaching of Grade 10 quadratic function. / Mathematics Education / M. Ed. (Mathematics Education)

Les données d’expression génique sont reconnues comme étant de grande dimension, etnécessitant l’emploi de méthodes statistiques adaptées. Mais dans le contexte des essaisvaccinaux, d’autres mesures, comme par exemple les mesures de cytométrie en flux, sontégalement de grande dimension. De plus, ces données sont souvent mesurées de manièrelongitudinale. Ce travail est bâti sur l’idée que l’utilisation d’un maximum d’informationdisponible, en modélisant les connaissances a priori ainsi qu’en intégrant l’ensembledes différentes données disponibles, améliore l’inférence et l’interprétabilité des résultatsd’analyses statistiques en grande dimension. Tout d’abord, nous présentons une méthoded’analyse par groupe de gènes pour des données d’expression génique longitudinales. Ensuite,nous décrivons deux analyses intégratives dans deux études vaccinales. La premièremet en évidence une sous-expression des voies biologiques d’inflammation chez les patientsayant un rebond viral moins élevé à la suite d’un vaccin thérapeutique contre le VIH. Ladeuxième étude identifie un groupe de gènes lié au métabolisme lipidique dont l’impactsur la réponse à un vaccin contre la grippe semble régulé par la testostérone, et donc liéau sexe. Enfin, nous introduisons un nouveau modèle de mélange de distributions skew t àprocessus de Dirichlet pour l’identification de populations cellulaires à partir de donnéesde cytométrie en flux disponible notamment dans les essais vaccinaux. En outre, nousproposons une stratégie d’approximation séquentielle de la partition a posteriori dans lecas de mesures répétées. Ainsi, la reconnaissance automatique des populations cellulairespourrait permettre à la fois une avancée pratique pour le quotidien des immunologistesainsi qu’une interprétation plus précise des résultats d’expression génique après la priseen compte de l’ensemble des populations cellulaires. / Gene expression data is recognized as high-dimensional data that needs specific statisticaltools for its analysis. But in the context of vaccine trials, other measures, such asflow-cytometry measurements are also high-dimensional. In addition, such measurementsare often repeated over time. This work is built on the idea that using the maximum ofavailable information, by modeling prior knowledge and integrating all data at hand, willimprove the inference and the interpretation of biological results from high-dimensionaldata. First, we present an original methodological development, Time-course Gene SetAnalysis (TcGSA), for the analysis of longitudinal gene expression data, taking into accountprior biological knowledge in the form of predefined gene sets. Second, we describetwo integrative analyses of two different vaccine studies. The first study reveals lowerexpression of inflammatory pathways consistently associated with lower viral rebound followinga HIV therapeutic vaccine. The second study highlights the role of a testosteronemediated group of genes linked to lipid metabolism in sex differences in immunologicalresponse to a flu vaccine. Finally, we introduce a new model-based clustering approach forthe automated treatment of cell populations from flow-cytometry data, namely a Dirichletprocess mixture of skew t-distributions, with a sequential posterior approximation strategyfor dealing with repeated measurements. Hence, the automatic recognition of thecell populations could allow a practical improvement of the daily work of immunologistsas well as a better interpretation of gene expression data after taking into account thefrequency of all cell populations.

With the realization that traditional instructional approach has not yielded satisfactory results, quasi-experimental and descriptive research designs were employed to investigate whether the application of constructivist instructional approach in the learning of stationary points in differential calculus by Grade 12 learners in South Africa would improve conceptual learning. Three Gauteng high schools of 204 Grade 12 learners constituted the research fields – one served as the control group while the other two represented the experimental group. Being a mixed-method research, quantitative data were gathered through pre-test and post-test while qualitative data were collected from classroom observations. Both inferential and descriptive statistical methods of data collection and analysis were used. The results obtained indicate that the experimental group demonstrated a better understanding of the concept of stationary points than the control group. / Mathematics Education / M. Sc. (Mathematics Education)

Text in Afrikaans / Die doel van die ondersoek sentreer rondom die verband tussen denkontwikkeling (‘n kognitiewe veranderlike) en prestasie in Afrikaans as tweede taal. Denkont-wikkeling is egter nie die enigste kognitiewe veranderlike wat prestasie in ‘n taal kan beïnvloed nie. Verbale begrip en geheue is ook as vername kognitiewe verander-likes geïdentifiseer. Affektiewe veranderlikes soos motivering, selfkonsep en angs, kan ook met prestasie in Afrikaans as tweede taal in verband gebring word. ‘n Empiriese ondersoek is uitgevoer waarby 174 hoërskoolleerders betrek is. Kognitiewe en affektiewe veranderlikes, asook leerstyl, is gemeet. Uit die empiriese ondersoek blyk dit dat selfkonsep, geheue, verbale begrip en motivering as die vernaamste veranderlikes beskou kan word wat met prestasie in Afrikaans as tweede taal verband hou. Denkontwikkeling is nie as ‘n vername faktor geïdentifiseer nie. Die bevindinge in die literatuurstudie en die empiriese ondersoek is bespreek om ouers en onderwysers van riglyne te voorsien om prestasie in Afrikaans te verhoog. / The aim of the research focusses on the relationship between thought development (a cognitive variable) and achievement in Afrikaans as a second language. Besides thought development, verbal understanding and memory recollection are also identified as distinctive cognitive variables. Affective variables such as motivation, self-concept and anxiety could also relate to performance in Afrikaans as a second language. Empirical research was conducted amongst 174 High School learners. Cognitive variables, affective variables and learning styles were measured. Resulting from the empirical research, it appears that self-concept, memory recollection, verbal comprehension and motivation are the main variables that impact on the performance of Afrikaans as a second language. Thought process development was not identified as a main factor. Results from the literature study and the empirical research are discussed in order to assist parents and teachers with guidelines to increase achievement in Afrikaans as a second language. / Psychology of Education / M. Ed. (Sielkundige Opvoedkunde)